This makes it quite difficult to identify with cer- 

 tainty a cell of the type shown in figure 270, 

 where the cytoplasm is slightly grayish, and the 

 nucleus is shrunken and has started toward de- 

 generation. These are all features character- 

 istic of the primary erythrocyte. 



By the end of the second week of incubation 

 and from then on until hatching, the mature 

 erythrocytes are indistinguishal)le from defini- 

 tive cells (figs. 228, 229^271, and 272). The 

 three cells showing artifacts (figs. 273-275) 

 have already been mentioned and they illustrate 

 the defects in greater detail than do the low- 

 power drawings (figs. 227-229). Only the 

 cytosome is affected, never the nucleus. In the 

 cytoplasm any odd effect may be produced. 



Embryo thrombocytes 



The tlirombocytes of the embryo appear as a 

 definite cell line soon after the primary eryth- 

 roblasts are well established and are clearly 

 present by 68 hours incubation (fig. 225), but 

 whether they are present at 48 hours also (fig. 

 224) is still undetermined. The cells of 224 

 A, 2 and 224 B, 1 and 2 may be primary throm- 

 boblasts. When seen in its entirety, cell 224 B, 

 2 resembled in some respects a yolk sac macro- 

 phage such as shown in figure 308. The eryth- 

 rocytes contain hemoglobin and the thrombo- 

 cytes do not; so if the two cell types are mingled 

 they could be sorted rather readily by Ralph's 

 benzidine method. 



There is a peculiar behavior of the primary 

 erythrocytes at the 48-hour age that may have a 

 bearing on the problem. If the primary eryth- 

 rocytes taken up into the cannula are not dis- 

 charged immediately onto the slide and there 

 spread so that they dry quickly, they will clump 

 and degenerate, as the thrombocytes will at an 

 older age. Many cells will clump together and 

 their appearance when partly degenerated is 

 similar to figure 224 A, 11 and 12. If eryth- 

 rocytes and thrombocytes come from the same 

 primordial cell, perhaps erythrocytes for a short 

 period of early embryonic life assume a function 

 peculiar to thrombocytes. 



Sugiyama (1926) concluded that thrombo- 

 cytes of the early chick embryo first arose by a 

 transformation of megaloblasts into Unombo- 

 blasts. The transition involved loss of hemo- 



globin and reduction in cell size. Our studies 

 have not supported the idea that the thromjjocyte 

 series arises from hemoglobin-bearing cells, al- 

 though it is recognized that a close parallelism in 

 development exists between erythrocytes and 

 thrombocytes. Six other theories on the origin 

 of thrombocytes have been reviewed by Sugi- 

 yama. Some have called these small clumped 

 cells of the embryo "lymphocytes," and Sugi- 

 yama (1926) presents convincing evidence that 

 this is an error. Typical lymphocytes in his 

 preparations did not appear in the circulating 

 blood of the emijryo until 17 days' incubation. 

 Our observations agree with his and further show 

 that lymphocytes, even in the late embryo, are in- 

 constant, and that they do not Ijecome a constant, 

 calculable component of the blood initil after 

 hatching. 



It is ol)vious from the papers of Dantschakoff 

 (1908b) and Sugiyama (1926) that both au- 

 thors were studying the same cell, the embryonic 

 thrombocyte. Dantschakoff claimed that pri- 

 mordial cells (large lymphocytes) of the early 

 embryo, produced microblasts and then micro- 

 cytes (dwarf lymphocytes). She suggested that 

 the dwarf lymphocytes might be related to the 

 spindle cells (thromJjocytes) and that the dwarf 

 lymphocytes were different from the small lym- 

 phocytes that appeared rather late in embryonic 

 life. According to Dantschakoff (1916a) the 

 latter appeared between the fifteenth and seven- 

 teenth day of incidjation. 



By 65 hours of incubation, the thrombocyte 

 line is easily distinguishable from the erythro- 

 cyte line; the cells of the former tend to clump 

 readily but the latter no longer exhibit this prop- 

 erty, and since the erythrocyte line is a group of 

 cells fairly well synchronized in development, a 

 structural series does not logically lead back to 

 the blast cell such as shown in figure 225, 25: yet 

 a complete gradation of cells does lead from this 

 to the small embryonic thrombocytes ( fig. 225, 

 27), which supports the conclusion that the blast 

 cell shown in figure 225 Ijelongs to the thrombo- 

 cyte series. The same is true of the cell shown 

 in figure 276 from the same slide. Cell 25 in 

 figure 225 appears to be discharging cytoplas- 

 mic blebs; whereas, in figure 276 only large pro- 

 trusions are present. Both show cytoplasmic 

 spaces but pro])ably the feature that most strongly 

 suggests their thrombocytic affinity is the punc- 

 tate character of the nuclear chromatin. This 



130 



